"Limitations live only in our minds. But if we use our imaginations, our possibilities become limitless". Jamie Paolinetti
What it does
Based on McKinsey survey, 85% of the banks identify the new regulations in respect to capital allocation and liquidity as the main reason to push credit risk manager to broader role in the bank aligned with finance and treasury. Limitless is the FinTech App for Credit Limit Manager that oversees the development and monitoring of counterparty credit limits. It highlights opportunities to optimize the capital allocation and boost liquidity. .
How we built it
Limitless is built using FFDC Design Studio and angular.js, it empowered by insights from unsupervised and supervised ML models to guide Credit Manager decisions. The historic credit exposure is clustered using k-mean model and refreshed daily using the ML classification algorithms. It utilizes TCM Portfolio Optimization API to provide insights on counterparties. The credit exposure forecast for 2 weeks is done using the deep learning (LSTM). The data for ML training and validation is extracted from in-house client mangled DB. The client is using the core (Summit) credit limits management for the last 10 years.
Challenges we ran into
New technologies and skills
Accomplishments that we're proud of
We use FFDC Open API, FFDC Design Studio and angular.js; got meaningful results for ML models; able to leverage unsupervised, supervised and deep learning ML techniques.
We put together the end to end application workflow based on the user journey of Credit Manager persona.
What we learned
We learn about challenges faced by financial institutions in respect of credit risk, capital allocation and liquidity management; ML models and UXP insights
What's next for Limitless
Limitless won Best in City award in Bucharest city side HackForTheFuture. The next step is to compete for Global Hackathon for TCM.